COMPOSITES

SCIENCE AND
TECHNOLOGY
Composites Science and Technology 65 (2005) 1436–1446
www.elsevier.com/locate/compscitech

Functionalized composite structures for new generation airframes:

a review
a,*
Lin Ye , Ye Lu a, Zhongqing Su a, Guang Meng b

a
Laboratory of Smart Materials and Structures (LSMS), Centre for Advanced Materials Technology (CAMT), School of Aerospace,
Mechanical and Mechatronic Engineering (J07), The University of Sydney, NSW 2006, Australia
b
State Key Laboratory of Vibration, Shock & Noise, Shanghai Jiao Tong University, Shanghai 200030, PR China

Available online 19 January 2005

Abstract

The uncontroversial superiority of functionalized composite structures for new generation airframes has been well acknowledged
by the research community. Such an approach has the potential to substantially enhance system performance and reduce overall
manufacture–operation–maintenance expenditure. Recent progress in informatics and high-capability computing devices has offered
a brand-new springboard for the aerospace community to reshuffle its traditional R&D criteria for functionalized composite struc-
tures. Particularly, artificial intelligence (AI), an intriguing information processing technique, exhibits outstanding effectiveness in
accommodating the highly demanding requirements of new generation airframes. Appropriate utilization of AI techniques in func-
tionalized composite airframe design will contribute to the realization of high-capability intelligent systems.
The applications of advanced composite structures, artificial intelligence and sensing network techniques in aircraft industry are
briefly reviewed in this paper, in correlation with various novel concepts. As a specific case study, an AI technique-based composite
structure with the capability of structural health monitoring was developed. An artificial neural network was customized and trained
using digitized spectrographic characteristics extracted from a multi-point sensing network. The system was then validated by exe-
cuting on-line health diagnostics, and the results indicate excellent performance of AI techniques in functionalized composite
structures.

2004 Elsevier Ltd. All rights reserved.

Keyword: Composite structure

1. Introduction cern since they were introduced, and have exponentially

A challenging issue presented in the aircraft commu-
nity is how to substantially enhance the performance of
engaged materials and structures so as to strengthen
their integrity and reliability, and meantime lower man-
ufacture–operation–maintenance cost. Owing to excel-
lent and unique mechanical properties, composite
structures have offered a promising solution to this con-
*
Corresponding author. Tel.: +61 2 9351 4798; fax: +61 2 9351
3760.
E-mail address: ye@aeromech.usyd.edu.au (L. Ye).
pervaded into various commercial applications, serving
as an attractive candidate to realize current and future
trends of new generation aero-vehicles [1,2].
Biomimetics, developed since last decade, has func-
tioned as a brand-new springboard for engineering
communities to reshuffle their traditional R&D princi-
ples. Biomimetics literally means imitation of life, and
more practically this concept refers to an interdisci-
plinary effort aimed at understanding biological princi-
ples and then applying them to improvement of
existing technologies or creation of entirely new tech-
nologies [3]. As one realization of biomimetics, artifi-

0266-3538/$ - see front matter
2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compscitech.2004.12.015
 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446
cial intelligence (AI), a novel information processing
technique emerged in the mid-1950s, is reputably
capable of learning, perception, and ratiocination.
Such a technique can be used to deal with a wide
variety of complex scientific and engineering problems.
At present, applications of AI technique in the aircraft
industry range from material design, manufacture, air-
craft maneuver, combat training, to traffic control and
damage detection [4].
Combination of the two above-mentioned intriguing
technologies stimulates the concept of AI-inspirited
functionalized composite structures. It is anticipated,
via such a principle, to achieve essential improvement
in integrity, reliability, performance, testability,
robustness and cost effectiveness of new generation
aircraft.
The status of on-going research and applications of
advanced composite structures and AI techniques, as
well as sensing networks, in aircraft industry is briefly
reviewed in this paper. As a specific application, a func-
tionalized carbon fibre/epoxy composite structure fea-
turing AI-supported structural health monitoring
(SHM) capability is introduced with validation. In this
system, an artificial neural network (ANN) is designed
and trained with signal spectrographic characteristics
extracted from a multi-point sensing network.
2. Composite materials and structures in airframes
2.1. Historical retrospect
Since patented by the DeHavilland Aircraft Co. in
the 1930s [5] and probably first introduced into commer-
Table 1
Application of composite materials in some typical aircraft [24]
Aircraft (military) Composite components
C-17 Large tail cone, winglet skins, fairing, etc.
AV-8B Wing trailing edge, wing skin, flap,
aileron, etc.
F-15 Horizontal and vertical tail skin, speed
brake, etc.
F-16 Horizontal and vertical tail skin, control
surface, etc.
F-22 Wing, fuselage, empennage, forward-
fuselage frame, etc.
F/A-18 E/F Fuselage skin, wing skin, etc.
Eurofighter aircraft (EFA) Monocoque, cured frames, longerons,
wetted area, etc.
Advanced jet fighter (AJF) Vertical tail stabilizer, etc.
Joint strike fighter (JSF-F35) Fuselage skin, upper wing skin tails,
control surface, inlet duct, engine access
cover panel, etc.
AH-64A Lower leading-edge fairings, stabilator,
etc.
1437
cial use as fuselage skin for Vultee BT-15 trainer plane
in 1944 [6], composite materials have been increasingly
employed by aeronautic and astronautic industries.
Based on matrix materials, the composites employed
in aircraft can be roughly categorized as polymer–ma-
trix, cement–matrix, metal–matrix, carbon–matrix and
ceramic–matrix composites [7], where the material
embedded in matrix behaves as the key structural prop-
erties, while the matrix serves as the binder to shape a
structural entity. Due to the superb strength-to-weight
and stiffness-to-weight ratios, substantial weight saving
and performance improvement can be achieved through
the utilization of composite materials compared with
conventional counterparts such as aero-aluminium alloy
[6]. For instance, 40–60% weight reduction is expected
by using high-strength titanium alloys and metal matrix
composites (MMCs) with oriented continuous fibre
reinforcement for early-21st-century aircraft [8]. Allied
to advantages in mechanical properties, their excellent
thermal, electric and magnetic features are also favor-
able for aircraft flutter suppression [9], fatigue-crack
reinforcement and repair [10,11], debris prevention
[12,13], temperature resistance [14], and damage detec-
tion [15–20]. By way of illustration, applications of com-
posite materials in some typical commercial and military
aircrafts are summarized in Table 1 [24], while detailed
reviews of their applications can be referred to elsewhere
[21–23].
2.2. Future visions of boeing and airbus
Targeting safety-essence, cost-efficiency, airworthi-
ness, system integrity, commonality and environmental
compatibility, major aircraft manufacturers are now
Aircraft (commercial) Composite components
Boeing 737 Aileron, elevator, rudder control surface,
etc.
Boeing 747 Fairings, nacelle components, winglet,
etc.
Boeing 757 Control surface, landing gear door, thrust
reverser, block doors, etc.
Boeing 767 Empennage, exterior surfaces, horizontal
stabilizer, etc.
Boeing 777 Rudder, fuselage side panel, stabilizer,
control surface, etc.
Airbus 300 Radome, pylon fairings, wing cover
panels, etc.
Airbus 310 Pylon fairings, fin leading, rudder, apron,
spoiler, etc.
Airbus 320 Stabilizer, fin, horizontal tailplane, etc.
Airbus 330 Flaps, empennage, floor beam, pressure
bulkhead, control lever, etc.
Airbus 340 Flaps, empennage, cockpit furnishings,
control lever, etc.
1438 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446
actively involved in the research and development of
innovative functionalized materials and structures for
the next generation aero-vehicles.
As the biggest commercial aircraft supplier world-
wide, Boeing Co. recently publicized its new generation
super-efficient commercial aircraft Dreamliner B-7E7,
which will allow future airlines to transport passengers
with the most economical operation cost. Competitively,
Airbus also launched its ambitious program, the A380
project, in 2000. This jumbo is claimed to be the most
advanced, spacious and efficient transportation vehicle
that human beings can conceive.
These two masterpieces are being individually devel-
oped, based on BoeingÕs and AirbusÕs divergent antici-
pations of future commercial transportation operation,
reflecting their unique visions for the crucial market of
competition. However, one mutual feature of both jets
is the unprecedented consumption of composite materi-
als. Dreamliner B-7E7 will be the first commercial air-
craft in which the majority of its structures, including
fuselage and wings, will be made of lightweight, super-
strong blends of carbon fibres and epoxy [25,26], com-
pared with the utility of 11% in its latest jet B-777. Such
a measure contributes to at least 3% fuel saving. Under
the exclusive structural design criteria proposed by Air-
bus, as schematically shown in Fig. 1, 22% of the
components of A380 will be manufactured using com-
posites (Fig. 2 [27]). Additionally, A380 will be the first
large commercial aircraft with the CFRP composite cen-
tre wing box. It boasts a weight saving of up to one and
a half tonnes compared to other aircrafts using the most
advanced aluminum alloys.
Such massive consumption of composite materials
motivates innovations in material manufacturing. Air-
bus introduces the latest manufacturing technologies
Fig. 1. General design criteria for A380 fuselage and empennage [27].
for A380, including automated fibre placement (AFP),
automated tape laying (ATL), resin film infusion
(RFI) and resin transfer moulding (RTM). Based on
these, the novel concepts of GLARE and laser beam
welding (LBW) are established for the jumbo. Impreg-
nated with an epoxy adhesive, GLARE is a kind of hy-
brid material built up from alternating slight overlap
layers of aluminum foils with unidirectional glass fibres,
leading to a single or double curved skin to obtain local
reinforcement. It is estimated that the application of
GLARE will lead to remarkable improvement in resis-
tance to corrosion and fire. The adoption of LBW, by
replacing the traditional riveting process for stringers
of lower fuselage panels, is aimed at reducing crack
growth in the aircraft skin [27].
3. Artificial intelligence and its applications in aircraft
industry
Defined as the simulation of human intelligence so as
to efficiently use ÔknowledgeÕ at a given step toward solv-
ing a problem [28], AI serves as a powerful solution to
complex engineering problems, for which conventional
straightforward logical algorithms are usually ineffi-
cient. At this stage, several variants originating from
fundamental AI concept can be found in applications,
namely expert system (knowledge-based system), fuzzy
logic, inductive learning, genetic algorithms and artifi-
cial neural network.
In particular, expert system and neural network are
most efficient. Basically, an expert system contains the
knowledge database for solving a particular kind of prob-
lem by means of explicit ÔIF–THENÕ rules [4]. Though
promising, most expert systems are unwieldy in their re-
 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446 1439

Fig. 2. Utility of composite structures on A380: (a) monolithic CFRP and thermoplastics; (b) materials distribution (weight breakdown).

sponse to new stimulants that are not included in their
database. With a similar mechanism but beyond the capa-
bility of an expert system, artificial neural network
(ANN) can be trained using a series of typical inputs
and their corresponding expected outputs, to establish
an implicit non-linear and multi-dimensional correlation
between them while avoid exploring the constitutive rela-
tion for a complicated system. Inherently endowed with
talents in adaptability, robustness and parallelism, the
ANN technique has found substantial applications in
pattern recognition, classification, function approxima-
tion, signal processing and system identification.
More recently, AI has been accepted for developing
control systems for both military fighters and commer-
cial jetliners [29–31]. In such approaches, massive data
for various flight conditions, preliminarily obtained
from experiments, numerical simulations or actual tests,
are hosted in a database and used to train the ANNs.
These well trained networks are able to automatically
adjust flying parameters to cater for unpredicted envi-
ronments. Such an AI-based flight control system is also
termed as fault tolerant flight control system (FTFCS)
in some studies, since it is valid even when some sensors
and actuators fail or new conditions appear [32]. The
1440 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446
approach is distinctly beneficial to tactical guidance and
survivability analysis, capable of keeping the vehicle in a
perfect operating condition all the time [33,34]. One of
the most successful paradigms using this concept is the
unmanned air vehicle (UAV) with an inertial navigation
system (INS) [35,36].
Zhang and Friedrich [37] systematically reviewed the
applications of neural network technique for composite
materials and structures, from fatigue prediction, wear
simulation, to manufacturing process monitoring and
curing analysis. It is concluded that a well trained
ANN is expected to be very helpful in predicting mate-
rial properties before manufacturing/testing the actual
composites, and in simulating the relationships between
various manufacturing parameters and material perfor-
mance. In particular, these predicted results can be
used as a basis for a computer-based processing opti-
mization to design new composite materials with re-
duced need for experiment. Other relevant
applications of ANN include traffic control, transport
management, airport infrastructure [38–42] and com-
posite manufacturing [43–46].
4. AI technique-supported functionalized structures
4.1. Future vision
As one of the long-term goals for engineering
community, ageless aircraft with built-in intelligent
self-diagnosis, self-control and self-healing/adjustment/
compensation capabilities can serve as a solution to vari-
ous issues, such as operation safety and life-cycle cost [47].
In an inspiring vision, the research team in NASA has
put forward a novel Darwinian vehicle design concept
(survival of fittest aircraft) based on AI principles [48].
It is hypothesized that only those aircraft designs that
acclimatize best to commercialization (more evolvable,
more resilient and more adaptive) can survive. Within
this concept, aircraft design and manufacturing adap-
tively undergo modifications in terms of variable envi-
ronmental circumstances, thereby ensuring continuous
improvement in vehicle functionality yet without the
necessity of human intervention [49]. Under Darwinian
vehicle design concept, aircraft feature automatic config-
uration of wing shape and operation control, in which
new control algorithms, software and even hardware-
configurations can be remotely uploaded from central
control on ground to in-service aircraft via a wireless
network. It instantly adjusts on-board memory, band-
width, power, control features and flight software codes.
Parallely, updated flight information can also be real-
time downloaded from aircraft to central control, to up-
grade future design, which can be considered as spiral
and reciprocal interactions.
4.2. Self-structural health monitoring
More practically as an achievable short-term target in
relation to current efforts, airframes, primarily made of
advanced structures integrated with robust sensing sys-
tems and intelligent data-processing ability, have come
into existence, comprehensively syncretising self-health-
prognostics and self-rehabilitation capabilities with
structural functionalities.
Advanced structures in aero-vehicles can suffer from
abrupt external impacts or extension of internal defects
during service, potentially leading to catastrophic failure
of aircraft without timely detection and repair. Disas-
trous disassembly of China Airline CI611 flight (B747-
200B) on 25th May 2002 and Columbia space shuttle
tragedy on 1st February 2003 entailed 225 and 7 lives
lost, respectively. However, these events represented only
15% of total aviatic casualties in 2002 and 2003 [50].
Investigation showed that both catastrophes could be
attributed to continuous deterioration of initially non-
lethal damage under fatigue or thermal loadings [48,50].
Motivated by aviation safety, on-line SHM is a system
with the capability of detecting and interpreting adverse
ÔchangesÕ in a structure in real-time, so as to enhance reli-
ability and reduce life-cycle costs [51]. A reliable on-line
SHM system should possess automatic data acquisition
and processing, structural condition assessment, and
decision-making for corrective actions. However, as a
complicated non-linear inverse problem, SHM is difficult
to achieve with most existing logic methods [52].
Recent studies [53–59] have substantiated that the
ANN technique is an encouraging solution towards this
problem. In practice, various structural characteristics
associated with damage in the single time or frequency
domain, or static parameters, are normally employed
for network training due to their simplicity of capture
for damage identification. These parameters include
mode shape [55,56] and natural frequencies [53,57], or
combined modal information [60]; displacement [61],
acceleration spectra [62] or combined parameters of dis-
placement, velocity and acceleration [63]; applied force
[64], or static parameters such as strain [65,66], strain
history [67], auto-correlation function [68], impedance
[69], etc. Subsequent work [70] has also demonstrated
that a well-trained ANN exhibits considerable tolerance
and robustness for partially incomplete or noise-impaired
information. Such a characteristic is particularly ben-
eficial for the health surveillance of flying aircraft,
where environmental interference may severely affect
the acquired signal or where data acquisition cannot
be completely fulfilled.
4.3. Active sensor network
In practice, the acquisition of structural parameters
and the implementation of control/compensation are
 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446 1441

realized through sensors and actuators embedded in or space applications. A first step to adopt this technology
bonded on the structures. Rather than a complicated to prototype structures has been achieved by the devel-
sensing/actuating network, simple diagnostic sensors opment of special encapsulated patches for an adaptive
have already been widely adopted as standard equip- satellite mirror.
ment for most aircraft, to monitor working condition Attached on plate-like structures, such as skins of
parameters such as temperature, pressure and emissions. wing or vertical fin, piezoelectrics can be used to gener-
However, these sensors can perform only local data ate stress waves, i.e. Lamb waves. Such a plate wave can
acquisition and are effective for specific components or propagate over a relatively long distance, even in mate-
parameters only. As an innovation, Boeing asserts that rials with high attenuation ratio, such as CF/EP com-
a high performance sensing network involving diverse posites, and thus monitoring over a broad area can be
sensors will be installed on the Dreamliner B-7E7, capa- conducted with only a few transducers in a distributed
ble of automatically and continuously surveiling vehicle configuration, saving considerable testing time and
performance in real-time. Primary advantages of such a costs. Meanwhile, the entire thickness of the plate can
technology are cost-saving and significant extension of be interrogated in virtue of different propagation modes
service life. of Lamb waves, affording the ability to detect internal
Sensor or sensor network technology, an interdisci- damage as well as defects on skin surfaces [73].
plinary subject, is seen as an integrated element in the Though there is still a great distance from laboratory
overall development of functionalized structures, span- research to practical application, as a valuable step, two
ning areas of physics, chemistry, materials, molecular representative sensor network techniques, SMART
biology, fabrication, electronics and signal processing. Layer and HELP layer, are well acknowledged and
Basically, sensing devices for current and next genera- have demonstrated potential for a wide diversity of engi-
tion aircrafts must meet requirements of (1) endurance neering cases.
for general environmental injury, (2) long life for at least SMART Layer (also nominated as Stanford Multi-
5–10 years, and (3) simple and easy handling and attach- Actuator–Receiver Transduction Layer), developed by a
ment [71]. For higher performance, they should be char-
acterized by smaller size, lighter mass, higher sensitivity,
lower cost/power consumption, higher reliability and
quicker response to sudden changes, as well as easier
integration. To accommodate future aero-applications,
remote control and data transmission, toughness for
vibration and noise, little aging deterioration and less
wire or even wireless network layout are also preferred.
Over the past decades, smart sensor networks and
technologies have been well developed in relation to
MEMS, nanotechnology, microelectronics, communica-
tion networks and distributed computing, multi-func-
tional structures integrated with a built-in smart
sensing network, and relevant intelligent signal process-
ing software packages. Various sensing devices are used
for different applications. Piezoelectrics, optical fibre,
electro-rheological fluids, shape memory materials, and
magnetostrictive/electrostrictive materials are some typ-
ical examples in the field of aerospace applications.
Amongst these, piezoelectrics shows a particularly good
capacity to satisfy exigent applications, due to unique
mechanical strength, wide frequency response range,
and favorable costs. A number of piezoelectric films/
patches can be easily distributed yet with only minor ef-
fects on the structureÕs overall mechanical performance.
Furthermore, with the aid of advanced encapsulation
technology, adaptive thin patches with PZT wafers have
been manufactured [72]. This technique is characterized
by a wide adaptability concerning shape and materials
to meet the requirements of a great variety of applica-
tions. The intention of this approach is to prepare and Fig. 3. Commercialized sensing networks, (a) SMART Layer;
encourage the establishment of adaptive structures of (b) HELP layer.
1442 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446

research team in Stanford University and commercial- to central control unit

ized by Acellent Technologies, Inc. [74,75], integrates

a certain number of distributed piezoceramics into a central
control
dielectric film to configure a network, as shown in Two-Way Switcher circuit
Fig. 3(a). As an extra thin and flexible ply fabricated
by a printed circuit technique, a SMART Layer can PZT Sensor

172.5 mm
be either inserted into structure or bonded onto surface
to function as an active sensor network, without notice-
able degradation of host structural integrity. Two-Way Switcher

Different to SMART Layer that generates and mea- to CCIU

sures the stress wave, HELP layer (Hybrid Electromag- 172.5 mm

netic Performing Layer), in Fig. 3(b), developed by the
to central control unit
French Aeronautics and Space Research Center (ON-
ERA), is based on the interaction between electromag-
netic field and structural abnormality. Such a layer (a)
Valid Coverage Area
allows the detection of local variations of electric con-
ductivity and dielectric permittivity induced by damage.
Mechanical damages, as well as thermal defects, can be
diagnosed due to their evident influence on electromag-
netic properties [76].

5. Case study

As a case study for elucidation, a functionalized CF/

EP composite structure with artificial neural algorithm-
supported SHM capability, developed at the Laboratory
of Smart Materials & Structures (LSMS) of the Univer-
sity of Sydney, is here presented.
In practice, aircraft components can be in various
geometries. The traditional or even optimized trans-
ducer allocation design cannot expediently provide an
efficient solution, where the distribution of actuators/
sensors and signal excitation/acquisition must be indi-
vidually considered upon each application. An active
sensor network technique was accordingly established
to simplify the transducer network design. Four PZT
disks were collocated to encircle a square area of 172.5
mm · 172.5 mm and controlled by a signal excitation/
acquisition circuit, as illustrated in Fig. 4(a), defined as
a standard sensor unit (SSU). Serving as the basic com-
ponent in an active sensor network, the SSU is colloca-
ble. Based on this concept, diverse sensor networks can
be conveniently tailored by assembling a set of SSUs to
flexibly accommodate different geometries and bound-
ary conditions. The distributed multi-sensor architecture
has also been experimentally validated as effective in
enhancing the signal-to-noise ratio (SNR) [58], and
exhibiting better robustness and reliability in a noisy
environment. As an example, an active sensor network
involving certain SSUs for a non-regular geometric iden-
tity is configured and displayed in Fig. 4(b).
An Intelligent Signal Processing & Patterns Recogni-
tion (ISPPR) package, inspirited by an artificial neural
algorithm and wavelet transform technique, was devel-
oped. Previous studies [77,78] have revealed that struc-
(b)
Fig. 4. Active sensor network: (a) standard sensor unit (SSU); (b)
customized sensing network for a geometrically non-regular structure.
tural damages can be quantitatively characterized by
energy spectra of the Lamb wave in structures under
inspection, which intuitively illustrates damage-scattered
Lamb waves over a dual-parameter space (time-fre-
quency domain). A novel concept was introduced in this
package to facilitate data processing, where the ex-
tracted spectrographic characteristics for each actua-
tor–sensor path in the active sensor network were
digitized and stored, referred to as Ôdigital damage fin-
gerprintsÕ (DDF), uniquely for the actuator–sensor path
under current damage situation [79]. By such an ap-
proach, a set of DDF can be obtained via all concerned
actuator–sensor paths, exclusively for a single damage
case. A multi-layer feedforward artificial neural network
supervised by an error-backpropagation (BP) algorithm
was constructed, in Fig. 5, and trained using the DDF
for various damage cases. Via the neural network, the
correlation between a structural damage case and
DDF was mapped, readily for further prognostics and
diagnostics.
 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446 1443

1
n11 w1−2 1 n12 w13 1 n13
w 1−1 Σ T1 Σ T2 Σ T3 ov1
im1
b11 b12 b13 (ξ )
1 2
Characteristics
im2
n2 n 2 n23
in time domain Σ T1 Σ T2 Σ T3 ov2
im3 b1 . . b2 . . b3 . (ξ )

.
2
. . 2
. . 2
.
. . . . . .
Characteristics in
time-frequency n1j nk2 nn3
domain imm Σ T1 Σ T2 Σ T3 ovn
wm1 − j w2j− k wk3−n (φ )
b1j bk2 bn3

Input First Neural Second Neural Output Diagnostic

Layer Processing Layer Processing Layer Layer Results

Fig. 5. Artificial neural network for designed functional composite structure.

A quasi-isotropic CF/EP composite laminate in the

configuration of [45/ 45/0/90]s was selected as the
structure to host the SHM functionalities. During devel-
opment, a customized actuating/sensing network with
programmed control circuits was incorporated onto
the composite laminate, to perform monitoring with
the aid of ISPPR [79]. Validation of such a system was
conducted by on-line evaluating a through-hole damage
or delamination in composite laminates (475 mm ·
475 mm · 1.275 mm), as shown in Fig. 6. Damage

Fig. 7. Damage prediction results: (a) hole-type damage detection;

(b) delamination damage detection ( , damage patterns for ANN
Fig. 6. A CF/EP composite laminate for validation of the technique. training).
1444 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446
diagnostic results, graphically described in Fig. 7, indi-
cate that satisfactory damage assessment for the com-
posite structure has been achieved [70].
6. Concluding remarks and prospect
The research and application status for advanced
composite structures, AI technology, and sensor/sensing
network technique, as well as their mutual integration,
are concisely addressed in terms of new generation air-
frames. A functional composite system with built-in
SHM capacity is presented as a case study, and the
validation exhibits promising prospect of AI tech-
nique-inspirited functionalized composite structures for
practical applications.
Stringent requirements from new generation cost-
efficient aero-vehicles have imposed unprecedented
challenges to the community, and have dramatically
motivated research and development for novel and prac-
tical techniques. In particular, the introduction of bio-
technology-based techniques shows encouraging
prospects in this field and claims an essential role in the
new millennium. Multi-functional structural systems,
comprising advanced composites and incorporating bio-
technology, will have a great impact on performance
enhancement, cost reduction and life extension.
Acknowledgements
Y. Lu and Z. Su thank the support of International
Postgraduate Research Scholarship (IPRS) from the
Department of Education Science, and Training
(DEST), Australia, and International Postgraduate
Award (IPA) from the University of Sydney.
References
[1] Buhl H. Advanced aerospace materials. Berlin: Springer; 1992.
[2] Flower HM. High performance materials in aerospace. Lon-
don: Chapman & Hall; 1995.
[3] Mann S. Biomimetic materials chemistry. New York: VCH;
1996.
[4] Pham DT, Pham PTN. Artificial intelligence in engineering. Int J
Mach Tool Manuf 1999;39(6):937–49.
[5] Chimenti DE. Guided waves in plates and their use in
materials characterization. Appl Mech Rev 1997;50(5):
247–84.
[6] Hoskin BC, Baker AA. Composite materials for aircraft struc-
tures. New York: American Institute of Aeronautics and Astro-
nautics; 1986.
[7] Chung DDL. Composite materials: functional materials for
modern technologies. London: Springer; 2003.
[8] Immarigeon J-P, Holt RT, Koul AK, Zhao L, Wallace W,
Beddoes JC. Lightweight materials for aircraft applications.
Mater Charact 1995;35(1):41–67.
[9] Suleman A. Adaptive composites modelling and application in
panel flutter and noise suppression. Comput Struct 2000;76(1–3):
365–78.
[10] Baker A. Bonded composite repair of fatigue-cracked primary
aircraft structure. Compos Struct 1999;47(1–4):431–43.
[11] Chiu WK, Koh YL, Galea SC, Rajic N. Smart structure
application in bonded repairs. Compos Struct 2000;50(4):433–44.
[12] Riddle R, Lesuer D, Syn C, Gogolewski R, Cunningham B.
Application of metal laminates to aircraft structures: prediction of
penetration performance. Finite Elem Anal Des 1996;23(2–4):
173–92.
[13] Schonberg WP. Protecting spacecraft against orbital debris
impact damage using composite materials. Compos Part A: Appl
Sci Manuf 2000;31(8):869–78.
[14] Windhorst T, Blount G. Carbon–carbon composites: a summary
of recent developments and applications. Mater Des
1997;18(1):11–5.
[15] Giurgiutiu V, Zagrai AN, Bao JJ. Piezoelectric wafer embedded
active sensors for aging aircraft structural health monitoring.
Struct Health Monitor 2002;1(1):41–61.
[16] Mickens T, Schulz M, Sundaresan M, Ghoshal A, Naser AS,
Reichmeider R. Structural health monitoring of an aircraft joint.
Mech Syst Signal Process 2003;17(2):285–303.
[17] Worden K, Manson G, Allman D. Experimental validation of a
structural health monitoring methodology: part I. Novelty
detection on a laboratory structure. J Sound Vibrat
2003;259(2):323–43.
[18] Manson G, Worden K, Allman D. Experimental validation of a
structural health monitoring methodology: part II. Novelty
detection on a GNAT aircraft. J Sound Vibrat
2003;259(2):345–63.
[19] Manson G, Worden K, Allman D. Experimental validation of a
structural health monitoring methodology: part III. Damage
location on an aircraft wing. J Sound Vibrat 2003;259(2):365–85.
[20] Edmonds J, Hickman G. Damage detection and identification in
composite aircraft components. In: Aerospace conference pro-
ceedings, IEEE, 18–25 March 2000;6:263–69.
[21] Ohnabe H, Masaki S, Onozuka M, Miyahara K, Sasa T. Potential
application of ceramic matrix composites to aero-engine compo-
nents. Compos Part A: Appl Sci Manuf 1999;30(4):489–96.
[22] Mouritz AP, Bannister MK, Falzon PJ, Leong KH. Review of
applications for advanced three-dimensional fibre textile compos-
ites. Compos Part A: Appl Sci Manuf 1999;30(12):1445–61.
[23] Kaczmar JW, Pietrzak K, Wlosinski W. The production and
application of metal matrix composite materials. J Mater Process
Technol 2000;106(1–3):58–67.
[24] Schwartz MM. Composite materials. Upper Saddle River,
NJ: Prentice-Hall PTR; 1997.
[25] Talbot D. BoeingÕs flight for survival. Technol Rev, September,
2003.
[26] www.boeing.com.
[27] www.airbus.com.
[28] Konar A. Artificial intelligence and soft computing: behavioral
and cognitive modeling of the human brain. CRC Press LLC;
2000.
[29] He MY, Jiang HL, Wei J, Li Y, Yang XY, Li J. Neural network
modeling of flight control system. IEEE AES Syst Mag
1998:27–9.
[30] Rogers JL. A parallel approach to optimum actuator selection
with a genetic algorithm, In: Presented at AIAA guidance,
navigation, and control conference AIAA 2000-4484, 14–17
August 2000. Denver, CO, USA, 2000.
[31] Richards RA. Application of multiple artificial intelligence
techniques for an aircraft carrier landing decision support
tool. In: Presented at IEEE 2002 world congress on compu-
tational intelligence, 12–17 May 2002. Honolulu, HI, USA,
2002.
 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446
[32] Campa G, Fravolini ML, Napolitano M, Seanor B. Neural
networks-based sensor validation for the flight control system of a
B777 research model. In: Presented at american control confer-
ence, 8–10 May 2002. Anchorage, AK, USA, 2002.
[33] McManus JW, Goodrich KH. Artificial intelligence (AI) based
tactical guidance for fighter aircraft. In: Presented at AIAA
guidance, navigation, and control conference, 20–22 August 1990.
Portland, OR, USA, 1990.
[34] Masud ASM, Metcalf P, Hommertzheim D. A knowledge-based
model management system for aircraft survivability analysis. Eur
J Operat Res 1995;84(1):47–59.
[35] Buskey G, Wyeth G, Roberts J. Autonomous helicopter hover
using an artificial neural network. In: Presented at international
conference on robotics & automation (ICRA 2001), May 2001.
Seoul, Korea, 2001.
[36] Nakanishi H, Inoue K. Development of autonomous flight
control systems for unmanned helicopter by use of neural
networks. In: Proceedings of the 2002 international joint confer-
ence on neural networks, 12–17 May 2002;3:2626–31.
[37] Zhang Z, Friedrich K. Artificial neural networks applied to
polymer composites: a review. Compos Sci Technol 2003;63(14):
2029–44.
[38] Moudani WE, Mora-Camino F. A dynamic approach for aircraft
assignment and maintenance scheduling by airlines. J Air Trans-
port Manage 2000;6(4):233–7.
[39] Jo G-S, Jung J-J, Yang C-Y. Expert system for scheduling
in an airline gate allocation. Expert Syst Appl 1997;13(4):
275–82.
[40] Wangermann JP, Stengel RF. Principled negotiation between
intelligent agents: a model for air traffic management. Artif Intell
Eng 1998;12(3):177–87.
[41] Ceylan H, Tutumluer E, Barenberg EJ. Neural network modeling
of slabs under simultaneous aircraft and temperature loading. In:
Presented at fourteenth engineering mechanics conference. Amer-
ican Society of Civil Engineers, 21–24 May 2000. Austin, TX,
USA, 2000.
[42] Blaess C, Tsiampalidis C, Vallee J-C. An application of artificial
intelligence for the safety in the neighbourhood of airport
runways. In: Proceedings of the IEEE ICTAIÕ96 workshop on
artificial intelligence for aeronautics and space, 16 November
1996. Toulouse, France, 1996.
[43] Maguire JF, Miller MA, Venketesan S. Processing of soft matter
and composites: integration of material sensors with process
models and intelligent control algorithms. Eng Appl Artif Intell
1998;11(5):605–18.
[44] Sticklen J, Kamel A, Hawley M, Delong J. An artificial
intelligence-based design tool for thin film composite materials.
Appl Artif Intell 1992;6(6):303–13.
[45] Marx WJ, Schrage DP, Mavris DN. An application of artificial
intelligence for computer-aided design and manufacturing. In:
Presented at international conference on computational engineer-
ing science: supercomputing in multidisciplinary analysis and
design, 30 July–3 August 1995. Mauna Lani, HI, USA, 1995.
[46] Adams DB, Watson LT, Gurdal Z. Blending of composite panel
designs using genetic algorithms. In: Proceedings of the interna-
tional parallel and distributed processing symposium (IPDPSÕ02),
15–19 April 2002.
[47] Noor AK, Venneri SL, Paul DB, Hopkins MA. Structures
technology for future aerospace systems. Comput Struct
2000;74(5):507–19.
[48] 48.www.nasa.gov.
[49] Noor AK, Doyle RJ, Venneri SL. Autonomous, biologically
inspired systems for future space missions. Adv Eng Software
2000;31(7):473–80.
[50] Su Z. PhD dissertation. Elastic wave-based quantitative damage
identification technique using active sensor network. University of
Sydney, 2004.
1445
[51] Kessler SS, Spearing SM, Soutis C. Damage detection in
composite materials using Lamb wave methods. Smart Mater
Struct 2002;11(2):269–78.
[52] Xu YG, Liu GR, Wu ZP, Huang XM. Adaptive
multilayer perceptron networks for detection of cracks in
anisotropic laminated plates. Int J Solids Struct
2001;38(32–33):5625–45.
[53] Okafor AC, Chandrashekhara K, Jiang YP. Delamination
prediction in composite beams with built-in piezoelectric devices
using modal analysis and neural network. Smart Mater Struct
1996;5(3):338–47.
[54] Liang YC, Hwu C. On-line identification of holes/cracks in
composite structures. Smart Mater Struct 2001;10(4):599–609.
[55] Tsou P, Shen M-HH. Structural damage detection and identifi-
cation using neural networks. AIAA J 1994;32:176–83.
[56] Yun CB, Bahng EY. Substructural identification using neural
networks. Comput Struct 2000;77(1):41–52.
[57] Chaudhry Z, Ganino AJ. Damage detection using neural
networks: an initial experimental study on debonded beams. J
Intell Mater Syst Struct 1994;5:585–9.
[58] Su Z, Ye L. Artificial neural algorithm-based damage assessment
for laminated composite structures using distributed actuators/
sensors. In: Paipetis AS, editor. Proceedings of the 5th interna-
tional symposium on advanced composites, 5–7 May 2003. Corfu,
Greece, 2003.
[59] Su Z, Ye L. Lamb wave propagation-based damage identification
for quasi-isotropic CF/EP composite laminates using artificial
neural algorithm. Part I: methodology and database development.
J Intell Mater Syst Struct 2005 [in press].
[60] Chang CC, Chang TYP, Xu YG. Structural damage detection
using an iterative neural network. J Intell Mater Syst Struct
2000;11:32–42.
[61] Chen HM, Qi GZ, Yang JCS, Amini F. Neural network for
structural dynamic model identification. J Eng Mech
1995;121:1377–81.
[62] Wu X, Ghaboussi J, Garrett Jr JH. Use of neural networks in
detection of structural damage. Comput Struct 1992;42(4):649–59.
[63] Zu B, Wu Z. Neural-networks-based structural health monitoring
strategy with dynamic responses. In: Chang F-K, editor. Struc-
tural health monitoring. 3rd ed. Boca Raton: Technomic Pub-
lishing Co.; 2001. p. 1418–27.
[64] Rhim J, Lee SW. A neural network approach for damage
detection and identification of structures. Comput Mech
1995;16:437–43.
[65] Hwu CB, Liang YC. Crack identification by artificial neural
network. Key Eng Mater 1998;145–149:405–10.
[66] Hwu CB, Liang YC. Hole/crack identification by static strains
from multiple loading modes. AIAA J 2001;39(2):315–23.
[67] Maseras-Gutierrez MA, Staszewski WJ, Found MS, Worden K.
Detection of impacts in composite materials using piezoceramic
sensors and neural networks. Proc SPIE Smart Struct Mater
1998;3329:491–7.
[68] Zubaydi A, Haddara MR, Swamidas ASJ. Damage identification
in a shipÕs structure using neural networks. Ocean Eng
2002;29:1187–200.
[69] Lopes Jr V, Park G, Cudney HH, Inman DJ. Smart structures
health monitoring using artificial neural network. In: Chang F-K,
editor. Structural health monitoring. 2nd ed. Lancaster: Tech-
nomic Publishing Co.; 1999. p. 976–85.
[70] Su Z, Ye L. Lamb wave propagation-based damage identification
for quasi-isotropic CF/EP composite laminates using artificial
neural algorithm. Part II: implementation and validation. J Intell
Mater Syst Struct 2005 [in press].
[71] The Mathworks, Inc. Data acquisition toolbox (userÕs guide), Ver.
2.0., 2000.
[72] Rolfes R, Sickinger C, Wierach P, Wetzel A, Kärger L, Breitbach
E. Advances, current developments and future trends in thin-
1446 L. Ye et al. / Composites Science and Technology 65 (2005) 1436–1446
walled composite aerospace structural systems. In: Loughlan J,
editor. Proceedings of international workshop on recent advances
and future trends in thin-walled structures technology, 25, June
2004. Loughborugh, UK, 2004.
[73] Alleyne DN, Cawley P. The interaction of Lamb waves with
defects. IEEE Trans Ultrasonics, Ferroelectrics Frequency Con-
trol 1992;39(3):381–97.
[74] Lin M, Chang F-K. The manufacture of composite structures
with a built-in network of piezoceramics. Compos Sci Technol
2002;62(7–8):919–39.
[75] Boller C. Ways and options for aircraft structural health
management. Smart Mater Struct 2001;10(3):432–40.
[76] Lemistre MB, Placko D, Balageas DL. Evaluation of the
performances of the HELP layer SHM system using both
DPSM simulations and experiments. In: Chang F-K, editor.
Structural health monitoring. 3rd ed. Lancaster: DEStech Pub-
lications, Inc.; 2003. p. 903–10.
[77] Su Z, Ye L, Bu XZ. A damage identification technique for CF/EP
composite laminates using distributed piezoelectric transducers.
Compos Struct 2002;57(1–4):465–71.
[78] Su Z, Ye L. Fundamental Lamb mode-based delamination detec-
tion for CF/EP composite laminates using distributed piezoelec-
trics. Struct Health Monitor: An International Journal
2004;3:43–68.
[79] Su Z, Ye L. An intelligent signal processing and
patterns recognition technique for wave propagation-based
structural health monitoring. Smart Mater Struct 2004;13:
957–69.